1: % ****** Start of file apssamp.tex ******
2: %
3: % This file is part of the APS files in the REVTeX 4 distribution.
4: % Version 4.0 of REVTeX, August 2001
5: %
6: % Copyright (c) 2001 The American Physical Society.
7: %
8: % See the REVTeX 4 README file for restrictions and more information.
9: %
10: % TeX'ing this file requires that you have AMS-LaTeX 2.0 installed
11: % as well as the rest of the prerequisites for REVTeX 4.0
12: %
13: % See the REVTeX 4 README file
14: % It also requires running BibTeX. The commands are as follows:
15: %
16: % 1) latex apssamp.tex
17: % 2) bibtex apssamp
18: % 3) latex apssamp.tex
19: % 4) latex apssamp.tex
20: %
21: \documentclass[twocolumn,showpacs,preprintnumbers,superscriptaddress,a4paper]{revtex4}
22: %\documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
23: %\documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
24:
25: % Some other (several out of many) possibilities
26: %\documentclass[preprint,aps]{revtex4}
27: %\documentclass[preprint,aps,draft]{revtex4}
28: %\documentclass[prb]{revtex4}% Physical Review B
29:
30: \usepackage{amsmath}
31: \usepackage{amssymb}
32: \usepackage{subfigure}
33: \usepackage{epsfig}
34: %\usepackage{graphicx}% Include figure files
35: %\usepackage{dcolumn}% Align table columns on decimal point
36: %\usepackage{bm}% bold math
37:
38: %\nofiles
39:
40: \begin{document}
41:
42: \preprint{APS/123-QED}
43:
44: \title{
45: Electronic and magnetic properties of substitutional Mn clusters in (Ga,Mn)As
46: }
47: \author{Hannes Raebiger}
48: \affiliation{
49: COMP/Laboratory of Physics,
50: Helsinki University of Technology, POB 1100, 02015 HUT, Finland.
51: }
52: \author{Andr\'es Ayuela}
53: \affiliation{Donostia International Physics Centre (DIPC),
54: POB 1072, 20018 San Sebastian, Spain}
55: \author{J. von Boehm}
56: \affiliation{
57: COMP/Laboratory of Physics,
58: Helsinki University of Technology, POB 1100, 02015 HUT, Finland.
59: }
60: %
61:
62: \date{\today}% It is always \today, today,
63: % but any date may be explicitly specified
64:
65: \begin{abstract}
66: The magnetization and hole distribution
67: of Mn clusters in (Ga,Mn)As are investigated by
68: all-electron total energy calculations using the projector augmented wave
69: method within the density-functional formalism. It is found that the
70: energetically most favorable clusters consist of Mn atoms surrounding one
71: center As atom. As the Mn cluster grows the hole band at the Fermi
72: level splits increasingly and the hole distribution
73: gets increasingly localized at the center As atom. The hole distribution
74: at large distances from the cluster does not depend significantly
75: on the cluster size. As a consequence, the spin-flip energy differences
76: of distant clusters are essentially independent of the cluster size.
77: The Curie temperature $T_C$ is estimated directly from these spin-flip energies
78: in the mean field approximation. When clusters are present estimated
79: $T_C$ values are around 250 K independent of Mn concentration
80: whereas for a uniform Mn distribution we estimate a $T_C$ of about 600 K.
81: \end{abstract}
82: \noindent {\it PACS \# \ }
83:
84: \pacs{75.50.Pp; 85.75.-d}% PACS, the Physics and Astronomy
85: % Classification Scheme.
86: %\keywords{Suggested keywords}%Use showkeys class option if keyword
87: %display desired
88: \maketitle
89:
90: \section{Introduction}
91:
92:
93: In the diluted magnetic semiconductor (Ga,Mn)As the Mn atoms
94: substituting Ga ones act as acceptors that simultaneously provide both
95: local magnetic moments and spin-polarized \textit{p}-\textit{d}-type
96: delocalized holes (according to a simple model one hole per Mn atom) which are
97: generally believed to mediate the ferromagnetic
98: coupling~\cite{ohno-ea-1996,matsukura-ea-1998,ohno-1998,ohno-1999,szczytko-ea-1999,beschoten-ea-1999,dietl-ea-2000,ohno-matsukura-2001,akinaga-ohno-2002,kepa-ea-2003,coey-sanvito-2004,dietl-2004}.
99: This unique feature where magnetism is clearly intertwined with
100: semiconductor properties makes (Ga,Mn)As a particularly attractive
101: material both for basic research and spintronics
102: applications. The Curie temperature ($T_C$) has been found to depend
103: directly on the delocalized hole concentration~\cite{satoh-ea-1997,ohno-1998,ohno-1999,potashnik-ea-2001,edmonds-ea-2002,yu-ea-2003,ku-ea-2003,sorensen-ea-2003}. It has also been found that annealing of the molecular beam epitaxially grown
104: (Ga,Mn)As removes Mn interstitial donors which increases delocalized hole
105: concentration and consequently $T_C$~\cite{yu-ea-2002,yu-ea-2003,edmonds-ea-2004}. The disorder of the Mn atoms either increases $T_C$ further
106: for lower Mn concentrations~\cite{berciu-bhatt-2001,chudnovskiy-pfannkuche-2002,kaminski-dassarma-2002,sandratskii-ea-2004}
107: or decreases $T_C$~\cite{xu-ea-2005}.
108: Another mechanism which affects hole concentration (decreasingly)
109: is the donation of electrons by the As antisite defects in the as-grown
110: (Ga,Mn)As~\cite{potashnik-ea-2001,grandidier-ea-2000,sanvito-hill-2001,korzhavyi-ea-2002,kudrnovsky-ea-2004,tuomisto-ea-2004}.
111:
112:
113: However, an additional mechanism which surely affects magnetic coupling and
114: the related hole distribution as well as $T_C$ is the clustering of Mn
115: atoms. Mn$_{\rm Ga}$-Mn$_{\rm Ga}$ dimer clusters (denoted henceforth by Mn$_2$ and
116: called ``dimer'') and Mn$_{\rm Ga}$ - interstitial Mn complexes have been found
117: in as-grown (Ga,Mn)As samples from cross-sectional scanning
118: tunneling microscopy images~\cite{sullivan-ea-2003}.
119: %However,
120: %the Mn$_{\rm Ga}$ - interstitial Mn complexes disappear during
121: %annealing~\cite{edmonds-ea-2004}.
122: %% change on 19.5.05 by hra
123: Some of these Mn$_{\rm Ga}$ - interstitial Mn complexes may disappear during
124: annealing~\cite{edmonds-ea-2004}, but complexes where the interstitial Mn
125: is bound to substitutional Mn clusters can be very stable~\cite{mahadevan-zunger-2003}.
126: %%%%
127: The presence of
128: Mn$_2$-clusters is natural because for a characteristic Mn concentration
129: of $x=$ 5 \% and assuming a random Mn distribution the probability that
130: a Mn atom belongs to a cluster is $1 - 0.95^{12} \approx 0.46$.
131: Structural evolution during post-growth annealing can lead to
132: further clustering~\cite{hayashi-ea-2001,potashnik-ea-2001,yu-ea-2002}
133: but no precipitation has been found~\cite{hayashi-ea-2001}.
134: Recently several {\it first-principles} calculations based on the
135: density-functional-theory have been performed
136: to study the Mn clustering and the related changes in
137: magnetism~\cite{schilfgaarde-mryasov-2001,rao-jena-2002,mahadevan-zunger-2003,raebiger-ea-2004,sandratskii-bruno-2004,raebiger-ea-2005,xu-ea-2005}.
138: The general outcome of these calculations may be summarized as follows.
139: Substitutional Mn clustering in (Ga,Mn)As is energetically favorable up to
140: Mn$_3$- or Mn$_4$-clusters~\cite{schilfgaarde-mryasov-2001,raebiger-ea-2005}.
141: Large magnetic moments are formed at the Mn clusters in
142: (Ga,Mn)As~\cite{raebiger-ea-2004,sandratskii-bruno-2004}
143: [see also Ref.~\cite{rao-jena-2002} for the case (Ga,Mn)N)].
144: At the same time the \textit{p}-\textit{d} hole distribution grows
145: significantly at the Mn clusters reducing the relative amount of delocalized
146: holes available for the long range
147: coupling of the magnetic moments of the Mn
148: clusters~\cite{raebiger-ea-2004,raebiger-ea-2005}.
149: %% addition on 19.5.05 by hra
150: Interstitial Mn is also found to form a stable complex
151: with Mn$_2$~\cite{mahadevan-zunger-2003}.
152: %%%%
153:
154:
155:
156: The aim of this Paper is to study theoretically the electronic and magnetic
157: properties of substitutional Mn clusters
158: in (Ga,Mn)As paying special attention to the hole mediated magnetic
159: coupling between the magnetic moments of the Mn clusters
160: using {\it first-principles} methods.
161: %% addition 19.5.05
162: We limit ourselves solely to substitutional Mn clusters;
163: interstitial Mn, As antisite or other possible point defects
164: are not considered although they may influence the physics
165: of (Ga,Mn)As~\cite{potashnik-ea-2001,grandidier-ea-2000,sanvito-hill-2001,korzhavyi-ea-2002,kudrnovsky-ea-2004,tuomisto-ea-2004,mahadevan-zunger-2003}.
166: %%
167: The Paper is organized as follows.
168: The computational methods are presented in Sec. II, the results for single
169: substitutional Mn impurity, Mn cluster formation and magnetic coupling
170: including discussion are presented in Sec. III, and the conclusions are
171: drawn in Sec. IV.
172:
173:
174: \section{Computational methods}
175:
176: Spin-polarized total energy supercell calculations based on the
177: density-functional-theory are performed for (Ga,Mn)As.
178: The projector augmented-wave method together with the
179: generalized gradient approximation (GGA-PW91) for exchange-correlation
180: as implemented in the VASP code is
181: employed~\cite{perdew-wang-1992,kresse-furthmuller-1996,kresse-joubert-1999}.
182: The projector
183: augmented wave method has the advantage that it is basically an all-electron
184: method but nevertheless almost as fast as the usual plane-wave pseudopotential
185: method.
186: %%This all-electron feature is important for (Ga,As)Mn which is magnetic.
187: %% rewriting by hra, 19.5.05
188: However, a slight limitation of the VASP implementation is the fact that the
189: core states are kept fixed. The projector augmented-wave
190: potentials
191: and the pseudopotentials for plane waves
192: provided in the VASP package
193: were
194: compared with an all-electron full-potential linearized
195: augmented-plane-wave calculation
196: by
197: performing calculations for MnAs (a crystal having the GaAs structure but
198: Mn substituting Ga).
199: %% 'method' changed to 'calculation' by hra
200: The projector augmented-wave calculation was found to agree
201: closely with the linearized augmented-plane-wave calculation but to differ both
202: quantitatively and qualitatively from the plane-wave pseudopotential
203: calculation~\cite{raebiger-ea-2004}.
204:
205: In our calculations
206: plane-waves up to the 275 eV cut-off value are included;
207: %% addition on 19.5.05
208: total energy in several test systems change less than 1 meV
209: as the cut-off value is increased from 275 to 300 eV or from 275 to 325 eV.
210: %%
211: Since
212: (Ga,Mn)As is a half-metal a dense k-mesh is needed.
213: %% addition on 19.5.05
214: The Brilloin zone integrations are done using the linear tetrahedron method
215: with Bloechl corrections~\cite{kresse-furthmuller-1996}.
216: %%
217: The Monkhorst-Pack
218: $(0.14$ \AA$^{-1})^3$ mesh including the $\Gamma$-point was found
219: sufficiently accurate for the Brillouin zone sampling~\cite{raebiger-ea-2004}
220: and is used in this study.
221: One or two Mn clusters (comprising of up to 5 Mn atoms substituting
222: Ga sites) are included in supercells of 64, 96 or 128 atoms,
223: corresponding to $2\times2\times2$, $2\times2\times3$ and $2\times2\times4$
224: cubic zinc-blende unit cells, respectively. Thus, the Mn concentration
225: in our calculations varies in the experimentally relevant range
226: [Mn] = 1.6 \ldots 7.8 \%.
227: Various Mn cluster distributions and different spin alignments
228: (e.g. $\uparrow\uparrow$ vs. $\uparrow\downarrow$) are studied.
229: Since relaxation effects were found negligible in the 64 atom
230: supercell calculations, only unrelaxed lattice positions with the fixed
231: experimental lattice constant $a =$ 5.65 {\AA} are used for larger supercells.
232: The orbital decomposition analysis is performed using the standard methods
233: implemented in the VASP code~\cite{kresse-furthmuller-1996}. The
234: self interaction error and the spin-orbit interactions are expected to
235: be small for (Ga,Mn)As~\cite{perdew-zunger-1981,filippetti-ea-2004,wierzbowska-ea-2005,dasilva-ea-2004} and thus neglected in our calculations.
236:
237:
238:
239:
240:
241: \section{Results and discussion}
242:
243:
244:
245: \subsection{Mn$_{\rm Ga}$ impurity in the dilute limit}
246:
247: The calculated density of states (DOS) for a uniform~\cite{footnote}
248: substitutional Mn
249: distribution in our dilute limit of $x = 1.6$ \% (or one substitutional Mn
250: atom in the 128 atom supercell) is shown in Fig.~\ref{fig:single} for
251: reference. The formation of the DOS may be followed using the simple
252: branching diagram of Fig.~\ref{fig:single} (a). The removal of a
253: Ga atom creates a vacancy (V$_{\rm Ga}$) acting as a shallow triple acceptor. The
254: six-fold degenerate \textit{p}-type $t_2$-level of the V$_{\rm Ga}$ is occupied
255: by three electrons~\cite{laasonen-ea-1992,scherz-scheffler-1992}.
256: When the Mn atom ($3d^54s^2$) is placed to the V$_{\rm Ga}$, the
257: V$_{\rm Ga}$ \textit{p} states and the Mn \textit{d} states hybridize
258: into the $t_{b\uparrow}$ \& $t_{b\downarrow}$ bonding states, the
259: $e_{\uparrow}$ \& $e_{\downarrow}$ states, and the $t_{a\uparrow}$
260: \& $t_{a\downarrow}$ antibonding states. The three electrons from the V$_{\rm Ga}$
261: and the seven electrons from the Mn atom occupy the
262: three $t_{b\uparrow}$ bonding states, the two $e_{\uparrow}$ states,
263: the three $t_{b\downarrow}$ bonding states, and two of the
264: three $t_{a\uparrow}$ antibonding states leaving one unoccupied
265: state (Fig.~\ref{fig:single} (a)). The bands and the DOS are subsequently
266: formed as shown in Fig.~\ref{fig:single} (b).
267: The most important feature of the DOS is the single hump in the majority
268: spin ($\uparrow$) channel at the Fermi level ($E_F$) which makes (Mn,Ga)As
269: a half-metal (Figs~\ref{fig:single} (b) and (c)).
270: As discussed above, the hump is formed mainly from the
271: $t_{a\uparrow}$ antibonding states (Fig.~\ref{fig:single} (a)).
272: Each Mn atom contributes one hole state
273: in the unoccupied part of the hump as well as the net magnetic moment of
274: 4$\mu_B$ [two $e_{\uparrow}$ and two $t_{a\uparrow}$ electrons, see
275: Fig.~\ref{fig:single} (a)] ($\mu_B$ is the Bohr magneton).
276: Our calculated DOS in Fig.~\ref{fig:single} agrees closely with several
277: other independent calculations~\cite{sato-katayama-yoshida-2001,sanvito-ea-2001,kulatov-ea-2002,sato-ea-2003,bergqvist-ea-2003,sandratskii-ea-2004,wierzbowska-ea-2005}.
278: %
279: \begin{figure*}[hbtp!]
280: \begin{center}
281: \epsfig{file=single-wide-7.eps,width=\linewidth}
282: \caption{The calculated density of states (DOS) for the Mn concentration
283: of $x = 1.6$ \%. (a) Branching diagram for a substitutional Mn atom.
284: The states of the substitutional Mn atom are formed via the hybridization
285: of the V$_{\rm Ga}$ $t_2$ and Mn $3d$ states. (b) DOS. $E_F$ denotes the
286: Fermi level. (c) The magnification of the DOS around $E_F$.
287: }
288: \label{fig:single}
289: \end{center}
290: \end{figure*}
291: %
292:
293: %
294: \begin{figure}[htbp!]
295: \begin{center}
296: %\subfigure[]{
297: \psfig{file=holesplot.ps,angle=-90,width=0.9\linewidth}
298: %}
299: \caption{The calculated As $p$ hole distribution. The average number
300: of $p$ holes per As site at each As coordination cell is given. The numbers are obtained
301: by integration from the orbital decomposition. Note that the orbital
302: decomposition as well as the choice of the integration volume are not
303: unambiguous. Nevertheless, the relative magnitudes are directly comparable.
304: }
305: \label{fig:hole-loc}
306: \end{center}
307: \end{figure}
308: %
309:
310:
311: The spin-polarized hole density of the hump consists mainly of the Mn
312: $d$ part localized around the Mn atoms and the As $p$ part that is
313: \textit{delocalized} around the As atoms. This delocalization
314: can be seen in the calculated As $p$ projection of the hole distribution
315: shown in Fig.~\ref{fig:hole-loc} (the average number of $p$ holes
316: at the As atoms belonging to each coordination shell measured from
317: the closest Mn site is given; for comparison the number
318: of holes at the Mn site is about 0.13).
319: The peaks at the distances of 2.4 and 6.2 {\AA} belong
320: to the nearest and third nearest As coordination shells.
321: Otherwise the $p$ hole distribution appears to be quite even.
322: The $p$ hole distribution does not change significantly when the Mn
323: concentration is changed (cf. the solid and dotted lines in
324: Fig.~\ref{fig:hole-loc} corresponding to the Mn concentrations of
325: 1.6 and 2.1 \%, respectively).
326:
327:
328:
329:
330: \subsection{Formation of Mn clusters}
331:
332: \begin{figure}[tbp!]
333: \begin{center}
334: %\subfigure[]{
335: \epsfig{file=supercells-5.eps,width=\linewidth}
336: %}
337: \caption{
338: Calculated Mn cluster configurations and total energies. The dark and gray
339: balls denote the Mn and As atoms, respectively. Separated components are placed
340: at the maximum distance available. All the black horizontal bars show the
341: (ferromagnetic) separation (or binding) energies; the corresponding values
342: are given in the units of meV in the black bars. The gray horizontal bars
343: with the arrow diagrams show the spin-flip energies; the corresponding values
344: are given in the units of meV inside the gray bars. All the black and gray
345: horizontal bars are drawn in the same scale.
346: }
347: \label{fig:megatable}
348: \end{center}
349: \end{figure}
350:
351:
352:
353: The energetically most important calculated Mn cluster
354: configurations are shown in Fig.~\ref{fig:megatable}. The supercells used are
355: chosen such that the Mn concentration remains the same $x =$ 6.3 \%. However,
356: in the case of five substitutional Mn atoms in the supercell a slightly higher
357: concentration of 7.8 \% is allowed due to the computational limitations.
358: For the cases of 2-4 Mn atoms in the supercell the configuration
359: where the Mn atoms share the \textit{same
360: neighbouring} As site is always found to be energetically most favorable
361: (see the upper left-hand corners in Figs~\ref{fig:megatable} (a) - (c)).
362: Also, it is energetically favorable to form one cluster from two separate
363: components in all cases shown in Figs~\ref{fig:megatable} (a) - (c).
364: In the case of five Mn atoms in the supercell the energetically
365: most favorable configuration is obtained by attaching the fifth Mn atom
366: to the stable tetramer [Fig.~\ref{fig:megatable} (d)].
367: The energies needed to separate single clusters into two components
368: (or the binding energies for the components) are shown graphically
369: in Fig.~\ref{fig:megatable} as black horizontal bars which contain
370: also the corresponding energy values in the units of meV.
371: In all these cases the ferromagnetic order is the stablest magnetic phase.
372: By using the above
373: binding energies we have calculated the heats of reaction for the cluster
374: formation and found that the optimal cluster sizes are
375: tetramers being slightly more favorable than the
376: trimers~\cite{raebiger-ea-2005}. Our calculation showing that clustering
377: is energetically favorable is in agreement
378: with the experimental result that Mn$_2$ clusters appear already in the
379: as-grown (Ga,Mn)As samples~\cite{sullivan-ea-2003}. Further clustering
380: is expected during post-growth annealing.
381:
382:
383:
384:
385:
386: \subsection{Effects of clustering}
387:
388: The dilute-limit majority-spin hump at the Fermi energy $E_F$ in
389: Fig.~\ref{fig:single} (c) is found to split when Mn clusters
390: are formed, and new narrower unoccupied bands appear in the gap.
391: This is shown for the Mn monomer, dimer, trimer, and tetramer systems
392: at the constant Mn concentration of $x = 6.3 \%$ in Fig.~\ref{fig:dostrends}.
393: At the same time the hole density grows at
394: the As atom which is situated in the center of the Mn cluster. This
395: was shown for the Mn dimer in Ref.~\cite{raebiger-ea-2004}
396: and a similar splitting was also found in Ref.~\cite{sandratskii-bruno-2004}.
397: The increasing localization at the center As atom is reflected
398: in the increasing separation and size of the
399: split-off part of the hump in Fig.~\ref{fig:dostrends}.
400: The As $p$-projection in the split-off part of the hole DOS is seen
401: to increase in relation to the Mn $d$-projection in Fig.~\ref{fig:dostrends}
402: and even to exceed the Mn $d$ projection in the cases of the Mn trimer and
403: tetramer [Figs~\ref{fig:dostrends} (c) and (d)].
404:
405:
406: The integrated numbers of the Mn $d$ holes at the cluster Mn atoms
407: and the As $p$ holes at the center As atom of the clusters are given in
408: Table~\ref{tab:hole-loc}. The results in Table~\ref{tab:hole-loc} clearly
409: show that the number of the As $p$ holes grows relatively faster than
410: the number of the Mn $d$ holes as the cluster size increases.
411: In the case of the pentamer,
412: the fifth Mn atom lies outside the tetramer and, in addition to
413: the center As atom inside the tetramer, there are two equivalent
414: As atoms between the fifth Mn atom and the tetramer
415: [Fig.~\ref{fig:megatable} (d)].
416: In this case the holes in the tetramer part remain intact
417: whereas the numbers of the Mn $d$ and As $p$ holes
418: are seen to grow at the additional fifth Mn atom and the two associated
419: As atoms, respectively (see the last line of Table~\ref{tab:hole-loc}).
420: We expect that an additional sixth Mn atom would cause further hole
421: growth at the new center-As-atom.
422: %
423: \begin{figure*}[hbtp!]
424: \begin{center}
425: \epsfig{file=dostrends2-5.eps,width=\linewidth}
426: \caption{Densities of states (DOSs) around the band gap (upper figures) and
427: the magnifications of the majority spin DOSs around the Fermi energy ($E_F$)
428: (lower figures). Figures (a), (b), (c), and (d) give the DOSs for the
429: Mn monomer, dimer, trimer, and tetramer, respectively for
430: the Mn density of $x = 6.3 \%$.
431: }
432: \label{fig:dostrends}
433: \end{center}
434: \end{figure*}
435:
436:
437:
438:
439:
440:
441: The As $p$ component of the spin-polarized hole density as a function of
442: the distance to the closest Mn atom is shown in Fig.~\ref{fig:clus-hole-loc}
443: for several different cluster sizes. At short distances $r < 7$ {\AA}
444: to the closest Mn atom of the cluster, the hole concentration
445: increases with increasing number of Mn atoms.
446: %% changed 20.5.05
447: However, for larger distances $r > 7$ {\AA} the hole concentrations
448: approach the same curve.
449: %%
450: Therefore,
451: the long-distance magnetic coupling between the Mn clusters
452: depends only on their mutual distance, not on the size of the cluster.
453:
454: \begin{table}
455: \caption{Holes inside the Mn cluster. The numbers of the
456: Mn $d$ holes at the cluster Mn atoms and the As $p$
457: holes at the As atom which is the center of the Mn
458: cluster are given. The numbers are obtained
459: by integration from the orbital decomposition. Although the orbital
460: decomposition as well as the choice of the integration volume are not
461: unambiguous the given average numbers of $p$ holes are directly comparable.
462: }
463: \label{tab:hole-loc}
464: \begin{tabular}{l c c c}
465: \hline
466: \hline
467: System & Mn $d$ holes & As $p$ holes & $x$ (\%) \\
468: \hline
469: single Mn & 0.13 & 0.03 & 6.3 \\
470: dimer & 0.17 & 0.11 & 6.3 \\
471: trimer & 0.19 & 0.19 & 6.3 \\
472: tetramer & 0.21 & 0.29 & 6.3 \\
473: pentamer & 0.21 ($\times$4) & 0.28 ($\times$1) & 7.8 \\
474: & 0.19 ($\times$1) & 0.10 ($\times$2) & \\
475: \hline
476: \hline
477: \end{tabular}
478: \end{table}
479:
480: %
481: \begin{figure}[htbp!]
482: \begin{center}
483: %\subfigure[]{
484: \psfig{file=clusholesplot.ps,angle=-90,width=0.9\linewidth}
485: %}
486: \caption{The calculated As $p$ hole distribution for different cluster sizes.
487: For further information see the caption of Fig.~\ref{fig:hole-loc}.}
488: \label{fig:clus-hole-loc}
489: \end{center}
490: \end{figure}
491: %
492:
493:
494:
495:
496:
497:
498:
499: In addition to the different ferromagnetic cluster configurations above
500: also various other magnetic alignments were considered. As mentioned above,
501: the ferromagnetic alignment is always energetically most stable.
502: The intra-cluster spin-flip energy is relatively high
503: for the Mn dimer: 211 meV~\cite{raebiger-ea-2004}
504: and significantly higher inside the larger clusters, typically 300 meV and
505: more. The spin flip energy of a distant component describes the stability
506: of the long range ferromagnetic state and may also be used
507: estimating $T_C$ as will be discussed shortly.
508: The calculated spin-flip energies $\Delta$ for the spin-flips of distant Mn
509: atoms or entire clusters are given in Table~\ref{tab:1} as well as in
510: Fig.~\ref{fig:megatable} as gray horizontal bars containing also the
511: corresponding energy values in the units of meV. The Mn concentrations
512: $x$ in Table~\ref{tab:1} are chosen such that the closest Mn-Mn distance
513: is constant (10.6 {\AA}) except for the first row where the distance is
514: 13.8 {\AA}. The calculated $\Delta$
515: values vary relatively little, between 57 and 81 meV.
516: %% additional clarification, 19.5.05
517: This variation in the $\Delta$ values exhibits no evident trend;
518: e.g. by looking at configurations with one cluster fixed,
519: an increase of the second cluster size leads to either increase or decrease
520: of the $\Delta$ value.
521: These fluctuations may be caused by e.g. directional
522: effects~\cite{sanyal-ea-2003,mahadevan-ea-2004}.
523: %%
524: This shows clearly
525: that new Mn atoms in the clusters promote the holes mainly to the center
526: As atom of the cluster and thus do not increase the delocalized hole
527: concentration mediating the ferromagnetic cluster-cluster coupling.
528:
529:
530:
531: The following expression based on the mean-field approximation
532: may be used to roughly estimate
533: $T_C$~\cite{sato-ea-2003,kurz-ea-2002,turek-ea-2003}
534: %
535: \begin{equation}
536: T_C = \frac{2}{3 k_B}\frac{\Delta}{N},
537: \label{eq:tc}
538: \end{equation}
539: %
540: where $\Delta$ is the energy difference between the spinglass and
541: ferromagnetic arrangements in the supercell, and $N$ denotes the number
542: of the magnetic particles in the supercell. However, we choose to
543: calculate $\Delta$ as the difference between the anti-parallel
544: and parallel arrangements which is known to be a good approximation
545: ~\cite{kurz-ea-2002,turek-ea-2003}.
546: We treat the Mn clusters as single magnetic
547: particles, and calculate $\Delta$ from the cluster-cluster
548: spin-flip energies. The number of Mn clusters
549: in the supercell is alważs fixed to $N = 2$.
550:
551:
552:
553: First, we consider the simple antiferromagnetic-ferromagnetic
554: case. Using Eq.(~\ref{eq:tc}) we estimate that for a uniform Mn
555: monomer distribution $T_C$ is 220 K for $x = 3.1 \%$
556: (Table~\ref{tab:1}) and 620 K for $x = 6.3 \%$~\cite{raebiger-ea-2004}.
557: %% rewriting by Juhani 20.5.05
558: The increase of $T_C$ from 220 to 620 K is mainly due to the decrease of the
559: Mn-Mn distance from 13.8 to 9.8 {\AA}.
560: If at $x=6.3 \%$ the Mn atoms were distributed randomly
561: at the Ga sites, about $1-0.937^{12} = 54 \%$ of the Mn atoms
562: would belong to clusters.
563: Therefore it is of some interest to compare the uniform Mn-dimer
564: distribution at $x=6.3\%$ (all Mn atoms belong to dimers)
565: with the uniform Mn monomer distribution at $x=6.3 \%$.
566: The dimerization reduces $T_C$ from 620 K to 313 K (Table~\ref{tab:1}).
567: A similar dramatic decrease of $T_C$ is found at $x=12.5\%$ from a uniform
568: Mn distribution to the corresponding Mn$_4$ cluster distribution
569: in Ref.~\cite{sandratskii-bruno-2004},
570: but at $x=6.3\%$ a slight increase in $T_C$ is found from
571: a uniform Mn distribution to the corresponding dimer distribution.
572: In Ref.~\cite{xu-ea-2005} it was concluded quite generally that
573: clustering decreases $T_C$, which is in agreement with our findings.
574:
575:
576:
577: We follow the same procedure as above to estimate $T_C$ from
578: Eq.(~\ref{eq:tc}) for further cluster distributions. The estimated $T_C$
579: values are given in Table~\ref{tab:1} where the Mn concentrations $x$ are
580: chosen such that the closest Mn-Mn distance is always 10.6 {\AA} (except
581: for the Mn$_1$+Mn$_1$ case). As noted above $\Delta$ does not depend
582: significantly on the Mn concentration and therefore neither does $T_C$:
583: the calculated $T_C$ values vary between 220 and 313 K (Table~\ref{tab:1}).
584: This may be related to the fact that the average As $p$ hole distribution at large
585: distances ($> 7$ \AA) is almost independent of the
586: number of Mn atoms in each cluster. Thus, as long as the
587: distance between clusters is kept constant, $T_C$ does not vary
588: significantly though the number of Mn atoms in the clusters is changed.
589:
590:
591: In the mean field approximation, $T_C$ is proportional to $\Delta/N$
592: (Eq.(~\ref{eq:tc})). During clustering, $\Delta$ may be expected to decrease
593: exponentially with respect to the growing cluster-cluster distance
594: while the number of clusters $N$ in the denominator decreases
595: much more slowly. Thus, as already noted above for the uniform dimer
596: distribution, clustering decreases $T_C$ significantly.
597:
598:
599:
600:
601: With the present high quality (Ga,Mn)As samples where the amount of
602: the harmful interstitial Mn and As$_{\rm Ga}$ defects has been minimized
603: one achieves $T_C$'s in the range
604: of 159-173 K~\cite{edmonds-ea-2004,wang-ea-2004}. The mean
605: field approximation used here may be expected to give an upper limit in
606: estimating $T_C$. The obtained estimates ranging between 220 and 313 K in
607: the case of clustering (Table~\ref{tab:1}) agree quite reasonably with
608: these best experimental values.
609: %%% new discussion, added on 19.5.05
610: This is consistent with the fact that
611: real samples contain clusters~\cite{sullivan-ea-2003}
612: which, as was shown above, can
613: reduce $T_C$ strongly.
614: %%even to less than half the estimated value of around 600 K
615: %%for a uniform sample (with periodic Mn monomer substitution and no clusters).
616: %%
617: %%$T_C$ has also been estimated within the coherent potential approximation
618: %%using the Korringa-Kohn-Rostocker method \cite{sato-ea-2003}
619: %%and the tight-binding linear-muffin-tin-orbital method \cite{kudrnovsky-ea-2003}.
620: %%
621: %%The coherent potential approximation calculations representing a fully random
622: %%Mn substitution give $T_C$ values around 300 K for the Mn concentration
623: %%of $x=6 \%$~\cite{sandratskii-ea-2004,kudrnovsky-ea-2003,sato-ea-2003},
624: %%which is less than our uniform value of 600 K.
625: %%This may well indicate that
626: %%calculations including both microscopic and large scale disorder
627: %%could lead to even lower $T_C$ values.
628:
629:
630: %%The difference may be attributed to the inclusion of disorder
631: %%in the coherent potential approximation calculations.
632: %%Although the $T_C$ values for the random Mn substitution are rather close to our
633: %%values from the cluster-cluster calculations,
634: %%the $T_C$ values cannot be directly compared.
635: %%Possibly
636:
637:
638:
639:
640:
641:
642:
643:
644:
645:
646: \begin{table}
647: \caption{Spin-flip total energy differences
648: between two substitutional Mn clusters.
649: $\Delta$ is the energy
650: difference between the system with one of the two clusters in anti-parallel
651: spin state (A) with respect to the system with all spins ferromagnetically
652: aligned (FM). The 128 atom supercell is used.
653: }
654: \label{tab:1}
655: \begin{tabular}{l c c c c c c}
656: \hline
657: \hline
658: System & FM & & A & $\Delta$ (meV) & $T_C$ (K) & $x$ (\%) \\
659: \hline
660: \hline
661: Mn$_1$+Mn$_1$ & $\uparrow\cdot\cdot\uparrow$ & & $\uparrow\cdot\cdot\downarrow$ & 57 & 220 &3.1 \\
662: \hline
663: Mn$_2$+Mn$_1$ & $\uparrow\uparrow\cdot\cdot\uparrow$ & & $\uparrow\uparrow\cdot\cdot\downarrow$ & 73 & 282 & 4.7 \\
664: \hline
665: Mn$_3$+Mn$_1$ & $\uparrow\uparrow\uparrow\cdot\cdot\uparrow$ & & $\uparrow\uparrow\uparrow\cdot\cdot\downarrow$ & 63 & 244 & 6.3 \\
666: Mn$_2$+Mn$_2$ & $\uparrow\uparrow\cdot\cdot\uparrow\uparrow$ & & $\uparrow\uparrow\cdot\cdot\downarrow\downarrow$ & 81 & 313 & 6.3 \\
667: \hline
668: Mn$_4$+Mn$_1$ & $\uparrow\uparrow\uparrow\uparrow\cdot\cdot\uparrow$ & & $\uparrow\uparrow\uparrow\uparrow\cdot\cdot\downarrow$ & 57 & 220 & 7.8 \\
669: Mn$_3$+Mn$_2$ & $\uparrow\uparrow\uparrow\cdot\cdot\uparrow\uparrow$ & & $\uparrow\uparrow\uparrow\cdot\cdot\downarrow\downarrow$ & 66 & 255 & 7.8 \\
670: \hline
671: \hline
672: \end{tabular}
673: \end{table}
674:
675:
676:
677:
678:
679:
680: \section{Conclusion}
681:
682: Substitutional Mn clustering in the diluted (Ga,Mn)As magnetic
683: semiconductor is studied by means of spin-polarized all-electron
684: density-functional calculations. The clustering is an important
685: factor in the typical Mn concentration range from 1 to 10 \% because
686: already in the case of purely random Mn distribution the probability
687: that a Mn atom belongs to a cluster varies correspondingly from 0.1 to 0.7.
688: Furthermore, our calculations show that cluster formation is energetically
689: favorable. The energetically most stable clusters are found to consist of
690: Mn atoms that surround symmetrically the center As atom. The spin-polarized
691: hole distribution is found to get increasingly localized at this
692: center As atom and the majority spin DOS hump
693: at the Fermi level to split
694: when the cluster size increases from Mn$_1$ to Mn$_4$.
695: At the same time the hole density outside the cluster
696: depends relatively little on the number of Mn atoms in the cluster,
697: especially at larger distances ($> 1.2 \times$ lattice constant).
698: This implies that the long range ferromagnetic coupling
699: between two Mn clusters depends relatively little on the number
700: of Mn atoms in the clusters. Our calculated spin-flip energies
701: confirm this expectation. Using these spin-flip energies and the mean
702: field approximation we estimate Curie temperatures around 250 K for the
703: various studied cluster distributions independently of the Mn concentration.
704: In contrast, we estimate for a uniform Mn monomer distribution at
705: $x =$ 6.3 \% a high value of 620 K. The best present (Ga,Mn)As
706: samples achieve Curie temperatures of 159-173 K. Therefore, our
707: estimated Curie temperature values are consistent with the fact that the (Ga,Mn)As
708: samples are hampered by the clustering effect.
709:
710:
711:
712: \section{Acknowledgments}
713:
714: This work has been supported by the Academy of Finland through the Center of
715: Excellence Program (2000-2005). A.A. acknowledges the financing of the Basque
716: Government by the ETORTEK program called NANOMAT.
717: The authors thank Acad. Prof. R. M. Nieminen,
718: Prof. K. Saarinen, Prof. M. J. Puska, Dr. M. Alava, Mr. F. Tuomisto and Mr.
719: T. Hynninen for many valuable discussions. We acknowledge the generous
720: computing resourches of the Center for Scientific Computing (CSC).
721:
722:
723:
724: \begin{references}
725: \bibliographystyle{plain}
726: \bibliographystyle{prsty}
727: %\bibliography{omaabbrev,paperbib}
728: %\bibliography{omaabbrev,silicide,GMR,dft}
729: %\bibliography{omaabbrev,paperbib}
730:
731:
732: \bibitem{ohno-ea-1996}
733: H.~Ohno, A.~Shen, F.~Matsukura, A.~Oiwa, A.~Endo, S.~Katsumoto, and
734: Y.~Iye, Appl. Phys. Lett. {\bf 69}, 363 (1996).
735:
736: \bibitem{matsukura-ea-1998}
737: F.~Matsukura, H.~Ohno, A.~Shen, and Y.~Sugawara, Phys. Rev. B {\bf 57},
738: R2037 (1998).
739:
740: \bibitem{ohno-1998}
741: H.~Ohno, Science {\bf 281}, 951 (1998).
742:
743: \bibitem{ohno-1999}
744: H.~Ohno, J. Magn. Magn. Mater. {\bf 200}, 110 (1999).
745:
746: \bibitem{szczytko-ea-1999}
747: J.~Szczytko, W.~Mac, A.~Twardowski, F.~Matsukura, and H.~Ohno,
748: Phys. Rev. B {\bf 59}, 12935 (1999).
749:
750: \bibitem{beschoten-ea-1999}
751: B.~Beschoten, P.~A.~Crowell, I.~Malajovich, D.~D.~Awschalom, F.~Matsukura,
752: A.~Shen, and H.~Ohno, Phys. Rev. Lett. {\bf 83}, 3073 (1999).
753:
754: \bibitem{dietl-ea-2000}
755: T.~Dietl, H.~Ohno, F.~Matsukura, J.~Cibert, and D.~Ferrand,
756: Science {\bf 287}, 1019 (2000).
757:
758: \bibitem{ohno-matsukura-2001}
759: H.~Ohno and F.~Matsukura, Solid State Commun. {\bf 117}, 179 (2001).
760:
761: \bibitem{akinaga-ohno-2002}
762: H.~Akinaga and H.~Ohno, IEEE Trans. Nanotechn. {\bf 1}, 19 (2002).
763:
764: \bibitem{kepa-ea-2003}
765: H.~Kepa, Le~Van~Khoi, C.~M.~Brown, M.~Sawicki, J.~K.~Furdyna,
766: T.~M.~Giebultowicz, and T.~Dietl, Phys. Rev. Lett. {\bf 91}, 087205 (2003).
767:
768: \bibitem{coey-sanvito-2004}
769: J.~M.~D.~Coey and S.~Sanvito, J. Phys. D: Appl. Phys.{\bf 37}, 988 (2004).
770:
771: \bibitem{dietl-2004}
772: T.~Dietl, in \textit{Proceedings of 27th International Conference on
773: Physics of Semiconductors, Flagstaff, Arizona, USA, July 2004}, edited
774: by J.~Mendez, AIP Proceedings, cond-mat/0408561 2004 (unpublished).
775:
776: \bibitem{satoh-ea-1997}
777: Y.~Satoh, N.~Inoue, Y.~Nishikawa, and J.~Yoshino,
778: in \textit{3rd Symposium on Physics and Application of Spin-Related Phenomena
779: in Semiconductors, Senadai, Japan, November 1997}, edited
780: by H.~Ohno, Y.~Oka, and J.~Yoshino, p. 23.
781:
782: \bibitem{potashnik-ea-2001}
783: S.~Potashnik, K.~Ku, S.~Chun, J.~Berry, N.~Samarth, and P.~Schiffer,
784: Appl. Phys. Lett. {\bf 79}, 1495 (2001).
785:
786: \bibitem{edmonds-ea-2002}
787: K.~Edmonds, K.~Wang, R.~Campion, A.~Neumann, N.~Farley, B.~Callagher, and
788: C.~Foxon, Appl. Phys. Lett. {\bf 81}, 4991 (2002).
789:
790: \bibitem{yu-ea-2003}
791: K.~M.~Yu, W.~Walukiewicz, T.~Wojtowicz, W.~L.~Lim, X.~Liu, U.~Bindley,
792: M.~Dobrowolska, and J.~K.~Furdyna, Phys. Rev. B {\bf 68}, 041308 (R) (2003).
793:
794: \bibitem{ku-ea-2003}
795: K.~C.~Ku, S.~J.~Potashnik, R.~F.~Wang, S.~H.~Chun, P.~Schiffer, N.~Samarth,
796: M.~J.~Seong, A.~Mascarenhas, E.~Johnston-Halperin, R.~C.~Myers, A.~C.~Gossard,
797: and D.~D.~Awschalom, Appl. Phys. Lett. {\bf 82}, 2302 (2003).
798:
799: \bibitem{sorensen-ea-2003}
800: B.~S.~S\o rensen, P.~E.~Lindelof, J.~Sadowski, R.~Mathieu, and P.~Svedlindh,
801: Appl. Phys. Lett. {\bf 82}, 2287 (2003).
802:
803: \bibitem{yu-ea-2002}
804: K.~M.~Yu, W.~Walukiewicz, T.~Wojtowicz, I.~Kuryliszyn, X.~Liu, Y.~Sasaki,
805: and J.~K.~Furdyna, Phys. Rev. B {\bf 65}, 201303 (R) (2002).
806:
807: \bibitem{edmonds-ea-2004}
808: K.~W.~Edmonds, P.~Boguslawski, K.~Y.~Wang, R.~P.~Campion, S.~N.~Novikov,
809: N.~R.~S.~Farley, B.~L.~Gallagher, C.~T.~Foxon, M.~Sawicki, T.~Dietl,
810: M.~Buongiorno~Nardelli, and J.~Bernholc, Phys. Rev. Lett. {\bf 92},
811: 037201 (2004).
812:
813: \bibitem{berciu-bhatt-2001}
814: M.~Berciu and R.~N.~Bhatt, Phys. Rev. Lett. {\bf 87}, 107203 (2001);
815: C.~Timm, F.~Sch\"afer, and F.~von~Oppen,
816: Phys. Rev. Lett. {\bf 90}, 029701 (2003);
817: M.~Berciu and R.~N.~Bhatt, Phys. Rev. Lett. {\bf 90}, 029702 (2003).
818:
819: \bibitem{chudnovskiy-pfannkuche-2002}
820: A.~L.~Chudnovskiy and D.~Pfannkuche, Phys. Rev. B {\bf 65}, 165216 (2002).
821:
822: \bibitem{kaminski-dassarma-2002}
823: A.~Kaminski and S.~Das~Sarma, Phys. Rev. Lett. {\bf 88}, 247202 (2002).
824:
825: \bibitem{sandratskii-ea-2004}
826: L.~M.~Sandratskii, P.~Bruno, and J.~Kudrnovsky, Phys. Rev. B {\bf 69},
827: 195203 (2004).
828:
829: \bibitem{xu-ea-2005}
830: J.~L.~Xu, M.~van.~Schilfgaarde, and G.~D.~Samolyuk, Phys. Rev. Lett. {\bf 94},
831: 097201 (2005).
832:
833: \bibitem{grandidier-ea-2000}
834: B.~Grandidier, J.~P.~Nys, C.~Delerue, D.~Sti\'evenard, Y.~Higo, and M.~Tanaka,
835: Appl. Phys. Lett. {\bf 77}, 4001 (2000).
836:
837: \bibitem{sanvito-hill-2001}
838: S.~Sanvito and N.~A.~Hill, Appl. Phys. Lett. {\bf 78}, 3493 (2001).
839:
840: \bibitem{korzhavyi-ea-2002}
841: P.~A.~Korzhavyi, I.~A.~Abrikosov, E.~A.~Smirnova, L.~Bergqvist, P.~Mohn,
842: R.~Mathieu, P.~Svedlindh, J.~Sadowski, E.~I.~Isaev, Yu.~Kh.~Vekilov,
843: and O.~Eriksson, Phys. Rev. Lett. {\bf 88}, 187202 (2002).
844:
845: \bibitem{kudrnovsky-ea-2004}
846: J.~Kudrnovsk\'y, I.~Turek, V.~Drchal, F.~M\'aca, P.~Weinberger,
847: and P.~Bruno, Phys. Rev. B {\bf 69}, 115208 (2004).
848:
849: \bibitem{tuomisto-ea-2004}
850: F.~Tuomisto, K.~Pennanen, K.~Saarinen and J.~Sadowski,
851: Phys. Rev. Lett. {\bf 93}, 055505 (2004).
852:
853: \bibitem{sullivan-ea-2003}
854: J.~M.~Sullivan, G.~I.~Boishin, L.~J.~Whitman, A.~T.~Hanbicki, B.~T.~Jonker,
855: and S.~C.~Erwin, Phys. Rev. B {\bf 68}, 235324 (2003).
856:
857: \bibitem{mahadevan-zunger-2003}
858: P.~Mahadevan and A.~Zunger, Phys. Rev. B {\bf 68}, 075202 (2003).
859:
860: \bibitem{hayashi-ea-2001}
861: T.~Hayashi, Y.~Hashimoto, S.~Katsumoto, and Y.~Iye, Appl. Phys. Lett.
862: {\bf 78}, 1691 (2001).
863:
864: \bibitem{schilfgaarde-mryasov-2001}
865: M.~van~Schilfgaarde and O.~N.~Mryasov, Phys. Rev. B {\bf 63}, 233205 (2001).
866:
867: \bibitem{rao-jena-2002}
868: B.~K.~Rao and P.~Jena, Phys. Rev. Lett. {\bf 89}, 185504 (2002).
869:
870: \bibitem{raebiger-ea-2004}
871: H.~Raebiger, A.~Ayuela and R.~M.~Nieminen, J. Phys.: Condens. Matter {\bf 16},
872: L457 (2004).
873:
874: \bibitem{sandratskii-bruno-2004}
875: L.~M.~Sandratskii and P.~Bruno,
876: J.Phys.: Condens. Matter {\bf 16}, L523 (2004).
877:
878: \bibitem{raebiger-ea-2005}
879: H.~Raebiger, A.~Ayuela, J.~von~Boehm, and R.~M.~Nieminen,
880: J. Magn. Magn. Mater. {\bf 290-291P2}, 1398 (2005).
881:
882: \bibitem{perdew-wang-1992}
883: J.~P.~Perdew and Y.~Wang, Phys. Rev. B {\bf 45}, 13244 (1992); and references
884: therein.
885:
886: \bibitem{kresse-furthmuller-1996}
887: G.~Kresse and J.~Furthm\"uller, Phys. Rev. B {\bf 54},
888: 11169 (1996); G.~Kresse and J.~Furthm\"uller, {\em VASP the Guide}
889: (Vienna University of Technology, Vienna, 1999)
890: [http://tph.tuwien.ac.at/$\sim$vasp/guide/vasp.html]
891:
892: \bibitem{kresse-joubert-1999}
893: G.~Kresse and D.~Joubert, Phys. Rev. B {\bf 59}, 1758 (1999).
894:
895: \bibitem{dasilva-ea-2004}
896: A.~J.~R.~da~Silva, A.~Fazzio, R.~R.~dos~Santos, and L.~E.~Oliveira,
897: J. Phys.: Condens. Matter {\bf 16}, 8243 (2004).
898:
899: \bibitem{perdew-zunger-1981}
900: J.~P.~Perdew and A.~Zunger, Phys. Rev. B {\bf 23}, 5048 (1981).
901:
902: \bibitem{filippetti-ea-2004}
903: A.~Filippetti, N.~A.~Spaldin, and S.~Sanvito,
904: Chem. Phys. {\bf 309}, 59 (2004).
905:
906: \bibitem{wierzbowska-ea-2005}
907: M.~Wierzbowska, D.~S\'anchez-Portal, and S.~Sanvito,
908: Phys. Rev. B. {\bf 70}, 235209 (2004).
909:
910: \bibitem{footnote}
911: In this Paper the uniform Mn distribution means a regular,
912: periodic Mn sublattice with no clustering, calculated within the supercell approach.
913: This does not correspond to the homogeneous disorder
914: in the coherent potential approximation calculations.
915:
916: \bibitem{laasonen-ea-1992}
917: K.~Laasonen, R.~M.~Nieminen, and M.~J.~Puska,
918: Phys. Rev. B {\bf 45}, 4122 (1992).
919:
920: \bibitem{scherz-scheffler-1992}
921: U.~Scherz and M.~Scheffler, in \textit{Imperfections in III/V Materials},
922: Semiconductors and Semimetals {\bf 38}, p. 1 (1992).
923:
924: \bibitem{sanvito-ea-2001}
925: S.~Sanvito, P.~Ordej\'on, and N.~A.~Hill,
926: Phys. Rev. B {\bf 63}, 165206 (2001).
927:
928: \bibitem{sato-katayama-yoshida-2001}
929: K.~Sato and H.~Katayama-Yoshida,
930: Jpn. J. Appl. Phys. {\bf 40}, L485 (2001).
931:
932: \bibitem{kulatov-ea-2002}
933: E.~Kulatov, H.~Nakayama, H.~Mariette, H.~Ohta, and Yu.~A.~Uspenskii,
934: Phys. Rev. B {\bf 66}, 045203 (2002).
935:
936: \bibitem{bergqvist-ea-2003}
937: L.~Bergqvist, P.~A.~Korzhavyi, B.~Sanyal, S.~Mirbt, I.~A.~Abrikosov,
938: L.~Nordstr\"om, E.~A.~Smirnova, P.~Mohn, P.~Svedlindh,
939: and O.~Eriksson, Phys. Rev. B {\bf 67}, 205201 (2003).
940:
941: \bibitem{sato-ea-2003}
942: K.~Sato, P.~H.~Dederichs, and H.~Katayama-Yoshida,
943: Europhys. Lett. {\bf 61}, 403 (2003).
944:
945: \bibitem{sanyal-ea-2003}
946: B.~Sanyal, O.~Bengone, and S.~Mirbt, Phys. Rev. B {\bf 68}, 205210 (2003).
947:
948: \bibitem{mahadevan-ea-2004}
949: P.~Mahadevan, A.~Zunger, and D.~D.~Sarma, Phys. Rev. Lett. {\bf 93}, 177201 (2004).
950:
951: \bibitem{kurz-ea-2002}
952: Ph.~Kurz, G.~Bihlmayer, and S.~Bl\"ugel, J. Phys.: Condens. Matter
953: {\bf 14}, 6353 (2002).
954:
955: \bibitem{turek-ea-2003}
956: I.~Turek, J.~Kudrnovsk\'y, G.~Bihlmayer, and S.~Bl\"ugel,
957: J. Phys.: Condens. Matter {\bf 15}, 2771 (2003).
958:
959: \bibitem{wang-ea-2004}
960: K.~Y.~Wang, R.~P.~Campion, K.~W.~Edmonds, M.~Sawicki, T.~Dietl, C.~T.~Foxon,
961: and B.~L.~Gallagher in
962: \textit{Proceedings of 27th International Conference on
963: Physics of Semiconductors, Flagstaff, Arizona, USA, July 2004}, edited
964: by J.~Mendez, AIP Proceedings,
965: cond-mat/0411475, 2004 (unpublished).
966:
967: %%\bibitem{kudrnovsky-ea-2003}
968: %%J.~Kudrnovsk\'y, I.~Turek, V.~Drchal, F.~M\'aca, J.~Ma\v sek, P.~Weinberger,
969: %%and P.~Bruno, J. Supercond. {\bf 16}, 119 (2003).
970: %%
971:
972: \end{references}
973:
974: % \end{multicols}
975:
976: \end{document}
977: